Build Daily

Tinley Park · June 27, 2026

Friday, June 19, 2026

mood: momentum

The machine is a commodity. The thing that knows how to make any part is the moat.

Shipped a big inspection feature to production, then took a hard turn into something physical: making our own gaskets on demand.

Shipped

  • Inspection Sections, in production — a tech can now draw a movable, named box on the canvas (Kitchen, Bar, Basement) and any unit inside it gets grouped automatically. Each section can carry its own separate estimate. Shipped the whole stack to production
  • Photos that actually load — added per-unit and per-visit photo capture, then chased down a real bug where pictures saved fine but showed up blank on Android. Rebuilt the way photos are served (short-lived signed links instead of auth headers the image loader was dropping) and they paint reliably now
  • Pinch-zoom and pan on the canvas — a long stainless line no longer fits on a phone screen, so the sketch surface now zooms and pans like a real drawing, with the math kept correct so doors still land in the right spot at any zoom
  • Gasket fabrication, proven on one profile — started building the system to make a gasket for any cooler door ourselves. Wrote a tool that traces our dimensioned profile drawings into geometry, then generates a 3D-printable fixture (and a metal-cutting file) to weld that exact profile. Proved it end-to-end on three different profile shapes
  • The real moat is digital tooling — the welder is a commodity anyone can buy. What no competitor has is a profile database wired to a fixture generator, so we can produce the tooling for any door from its drawing for the cost of filament. That's what's worth building

Notes

Today reframed a pile of separate projects — the field app, the profile database, the welder, the fixtures — into one system: order to spec to material to cut to weld to ship. The lesson buried in it is a buy-vs-build rule. Buy the commodity (the machine). Build the part nobody else has (the software that turns any profile into tooling on demand). Picking the wrong machine — the kind that locks you into its own proprietary molds — would have quietly thrown the whole advantage away.